A3: Incarceration

Introduction

With this dataset of prison and jail information from 1970 to 2018, I plan to observe the proportion of women in jail and in prison. I am planning to analyse the jail population of women (female_jail_pop) from more recent years, as the data is more accurate from 2015 on due to how the data is collected. For prison populations I will be analyzing female populations (female_prison_pop) but also female populations by race (aapi_female_prison_pop,black_female_prison_pop,latinx_female_prison_pop,native_female_prison_pop,other_race_female_prison_pop, and white_female_prison_pop). I will look at the proportion of females by race in the prison system by analyzing the female population variables and their race categories. I am hoping to develop some insight on female populations in prison, and how their race may affect the size of population.

Summary Information

When I first began looking at this data, I wanted to look at the data from a marginalized standpoint, which is why I chose to focus on women and women minorities. I first wanted to understand what percentage of the prison population I am looking at. For more accurate information I decided to look at the proportion of women in jail for the years 2015-2018, and the proportion of women in prison for the years 2015-2016. In 2017 and 2018 there is no data on prison populations. In these years they began counting populations throughout the whole year, whereas in other years they counted until June and estimate the rest, so I believe these counts to be more accurate.

For the years 2015-2018, the average proportion of women populating jails in the United States was 0.1432655. For 2015-2016, the average proportion of women populating prisons in the United States was 0.0915272. There proportions are considerably low. For jail population there is no information on women based on their race, but there is for prison populations. Out of the 0.0915272 proportion of women in prison populations, 2.071163^{-4} are Asian American Pacific Islander, 0.0128005 are Black, 0.0076854 are Latinx, 0.0071462 are Native American, 5.6051774^{-4} are of a other race besides those listed, and 0.0725697 are White.

You can also observe the difference of these women prison populations from before 2015 and 2015 and on. For example here is the difference for Black females: 0.0170128, 0.0128005, and for white females: 0.0514232, 0.0725697.

One of the most interesting to look at is Black females in prison populations. As you can see from the above data many of the races saw an increase from before 2015 to after, except for Black females. In fact, the proportion of Black females in prisons begins to drop starting in the early 2000’s. Here is the proportion of Black females in prison populations and admissions from 1970 to 2016. In 1970 the Black female prison population averaged 0.6956522 and in 2016 it averaged 21.3913043

All of these findings are from averaging all the counties listed in this dataset. These trends do not speak for each individual county but provide a more accurate average for the United States.

The Dataset

Who collected the data?

• This data was collected by two Bureau of Justice Statistics data collections the Census of Jails and the Annual Survey of Jails. The Vera Insitiute of Justice compiled this data to create the datasets.

How was the data collected or generated?

• The two Bureau of Justice Statistics data collections are conducted on a cycle every so many years. The Census of Jails is conducted every five to eight years since 1970 and covers all jails. The Annual Survey of Jails has been conducted in non-census years since 1982 and covers almost every one of the largest jails and about one third of all jails. The Vera Institute of Justice used these data collections to create the Incarceration Trends data tool in 2015 and in 2016-2017 received a grant to update the data tool with newly released data. In 2018, Vera receieved another grant to work on county-level data and merged it with original jail datasets to create the first national dataset comparing jail and prison incarcerations at the county level. They have updated the data in 2019 and 2020 to include recent data.

Why was the data collected?

• The dataset was created due to the demand for more incarceration information on the county level. Much incarceration data focuses on the state-level, which helps comprehend the broad trends of incarcerations. However, this county-level data provides a narrower view where causes and effects are easier to identify.

How many observations (rows) are in your data?

• There are 153811 observations in the county-level dataset and 207074 observations in the jurisdiction-level dataset.

How many features (columns) are in the data?

• There are 121 features in the county-level dataset and 29 features in the jurisdiction-level dataset.

What, if any, ethical questions or questions of power do you need to consider when working with this data?

• When working with this data, I am reminded that this data is collected by the Bureau of Justice Statistics, which is under the Justice Department in the United States government. The government has notorious been unkind to minorities in the United States. The BJS website lists reports that “measure the magnitude, nature, and impact of crime in the nation.” and it is curious if they care more about how crime affects other citizens rather than how the criminal justice system affects citizens who a subjected to it.

What are possible limitations or problems with this data? (at least 200 words)

• Possible limitations or problems with this data include the numerous blank cells due to missing data and the number of cells that are estimated due to a lack of data. Through 1983 to 1999, there was only data available on admissions and releases so the populations counts are all estimates for these years. Another example, total jail admissions are measured through the last week in June and then multiple to get an annual count, which could be inaccurate in varying degrees by county. Counting until the end of June is common for other variables as well, and it also depends on year, like for the total jail population they start counting to December in 2015 and 2016, but before that stopped in June. There are also varying records for different states, such as some states exclude admission and release dates that are less than one year apart for certain years. For an accurate analysis, the data will have to be narrowed down because looking at the data broadly may be very inaccurate.

Trends Over Time Chart

This chart shows the average female prison population by race from 1970 to 2016. I included this chart to show some anomalies that I found interesting that may warrant further investigation. Firstly, there is a notable increase after 1990 in almost every race category, and for some the increase is more drastic. For Black women the increase is the most significant as it surpasses white women by approximately and average of 10, even though the white population in the US is much larger than the Black population. Secondly, between 1990 and 2000 there is sharp decline in the average population of Black women in prisons, and it continues to decline intermittently after that, while the white women population increases. You can also see that Asian American Pacific Islander, Latinx, and Native American Women are not included in this chart until a few years before 1990. Other races seems to stay at 0 until that time, and then it increases and decrease significantly in the span of a decade. These trends are very interesting and I believe could be explained by further analysis into current events at the time and how the data was collected for these groups.

Variable Comparison Chart

In this chart, I took a more focused look on something that caught my attention in the last chart. The average population of white females rose above the average population of black females around the year 2000, and I was curious if that was reflective at the county level. Each point represents a county in the United States and whether there is a greater population of Black women or white women. Some trends I am noticing are that there usually is a greater population of white women, but it is not greater by much. Yet if you notice the light blue dots that trail off to the right, when there is a greater population of Black women the difference is significant. An also curious observation is the string of dark blue dots rising quickly on the left side of the graph. This can’t be seen in the graph, but if you look at the data it is the same county over the years: Maricopa County, AZ. The significant difference is what is most aligned with population ratios in the United States, but it is only represented in one county. I think further analysis into the county would be beneficial.

Map

Looking further into my analysis on Black females in the prison system, I have created this map that shows the proportion of women in the prison system who are Black. I included this chart to show which states have a large proportion of Black women, even though the Black population in the US has never exceeded 13% according to the US Census Bureau. I chose to focus on Black females because they are the only minority that exceeds white females in recent years, and I thought that was important to visualize considering the majority of the United States is white. Some patterns I have noticed are that the states with higher percentages of Black females tend to be in the South and East parts of the country, while many West states have lower percentages. I would be curious to compare the population of Black females in the state to that of Black females in prison, to see if there is a correlation. I am also curious about the states where there are no count of females by race, and what their numbers would’ve added to this graph.